Successful Versus Failed Transition From Controlled Ventilation to Pressure Support Ventilation in COVID-19 Patients: A Retrospective Cohort Study

Abstract

Objectives: In patients with COVID-19 respiratory failure, controlled mechanical ventilation (CMV) is often necessary during the acute phases of the disease. Weaning from CMV to pressure support ventilation (PSV) is a key objective when the patient's respiratory functions improve. Limited evidence exists regarding the factors predicting a successful transition to PSV and its impact on patient outcomes.

Design: Retrospective observational cohort study.

Setting: Twenty-four Italian ICUs from February 2020 to May 2020.

Patients: Mechanically ventilated ICU patients with COVID-19-induced respiratory failure.

Intervention: The transition period from CMV to PSV was evaluated. We defined it as "failure of assisted breathing" if the patient returned to CMV within the first 72 hours.

Measurements and main results: Of 1260 ICU patients screened, 514 were included. Three hundred fifty-seven patients successfully made the transition to PSV, while 157 failed. Pao₂/Fio₂ ratio before the transition emerged as an independent predictor of a successful shift (odds ratio 1.00; 95% CI, 0.99-1.00; p = 0.003). Patients in the success group displayed a better trend in Pao₂/Fio₂, Paco₂, plateau and peak pressure, and pH level. Subjects in the failure group exhibited higher ICU mortality (hazard ratio 2.08; 95% CI, 1.42-3.06; p < 0.001), an extended ICU length of stay (successful vs. failure 21 ± 14 vs. 27 ± 17 d; p < 0.001) and a longer duration of mechanical ventilation (19 ± 18 vs. 24 ± 17 d, p = 0.04).

Conclusions: Our study emphasizes that the Pao₂/Fio₂ ratio was the sole independent factor associated with a failed transition from CMV to PSV. The unsuccessful transition was associated with worse outcomes.

Keywords: COVID-19; acute respiratory distress syndrome; mechanical ventilation; respiratory effort; transition to pressure support ventilation.

Figure 1.

Timetable of the study. CMV = controlled mechanical ventilation, PSV = pressure support ventilation.

Figure 2.

Linear mixed model with patients as random effects: A, Pao2/Fio2; B, Paco2; C, peak inspiratory pressure; and D, plateau between the success and the failure group. The entire models for all the variables are inserted in the supplements, section “Models for Pao₂/Fio₂, Paco₂, peak inspiratory pressure, and plateau pressure.”

Figure 3.

A, Kaplan-Meier ICU mortality, hazard ratio (HR) between the success and the failure group. The adjusted model is in the supplements section “adjusted model for ICU survival.” B, Kaplan-Meier probability of breathing without assistance, hazard ratio between the success and the failure group.